Social democratic tax policy updated : A study of the Norwegian tax reform of 1992

(1)

Social democratic tax policy updated

A study of the Norwegian tax reform of 1992 Johan Christensen

Master’s thesis, Department of Political Science

UNIVERSITY OF OSLO

22.05.2009

(2)

Preface

People always advised me to pick a small, well-defined topic for the Master’s thesis. As will become obvious once you start reading my thesis, I didn’t listen very well. Instead, I chose to address a really big question. I’m glad I did. Working on this thesis is by far the most intellectually stimulating exercise I have ever undertaken. It has given me the opportunity to dig into a fascinating theoretical literature, speak to people with extraordinary knowledge, and most importantly, gain in-depth understanding of a complex issue.

In particular, the interviews with policy-makers and experts sparked my interest for the subject and gave the project a clearer direction. I would therefore like to thank all my informants for their contribution.

Moreover, I would like to thank those without whom this project would not have been feasible. My supervisor (fall 2008 and spring 2009), colleague and friend Bent Sofus Tranøy has been a great source of intellectual inspiration. Rather than pulling the brake, he has always stimulated my ambitions. My good friend Helge Renå has been a fantastic discussion partner throughout the process. The effort he has put into reading and commenting on my work is truly remarkable and will not be forgotten easily.

I am also grateful to my dad, who spent half a day fighting with an earlier draft and came back with excellent comments, and to Øivind Bratberg, who gave me good advice and supplied me with some very interesting literature.

Of course, I take full responsibility for all remaining errors, problems, flaws, shortcomings, inaccuracies and spelling mistakes.

Johan Christensen Oslo, 22.05.09

(6)

Chapter 1: Introduction

1.1 Topic and research questions

The tax reform that came into effect in 1992 represents the most wide-reaching reform of the tax system in Norwegian history. The reform lowered the tax rates for both individuals and businesses drastically, cut down a whole forest of deduction schemes, and established new governing principles for tax policy. This instance of fundamental¹ policy change is the subject of the present study.

Since the tax system pays for the welfare system, the 1992 tax reform is part and parcel of the modern transformation of Norway’s social democratic welfare state regime. Yet, our understanding of the outcome of the tax reform is underdeveloped. Did these policy changes represent the death of egalitarian social democratic tax policy or rather its resurrection?

Couched in more analytical terms, the first question posed by this study is: Did the Norwegian tax reform of 1992 imply a liberal policy shift or a rational updating of social democratic policy?

Within the literature on institutional change in advanced political economies, a central argument is that these substantive outcomes of change are somehow contingent upon the process of change. Wolfgang Streeck and Kathleen Thelen (2005) argue that liberalization is usually the product of small, incremental changes in policy. Conversely, one would expect that non-liberal reforms are most often the result of large changes in policy.

This gives rise to the suspicion that the large scale of the tax policy changes in 1992 shaped the outcome of the reform, i.e. whether it led to liberalization or rational updating of policy.

The second question explored by this study is thus: How was the outcome of the 1992 tax reform conditioned by the large scale of policy change?

This potential relationship raises yet another question: What made large-scale reform possible? The institutionalist literature emphasizes the difficulty of such authoritative policy change, due to the status-quo bias of political institutions. However, I suspect that the

1 To denote policy changes of large scale, I use the expressions ‘large-scale’, ‘fundamental’, ‘authoritative’,

‘large’, ‘deep-seated’, ‘wide-reaching’ and ‘profound’ inter-changeably.

(7)

consensual Norwegian political system represents an institutional setting that is more conducive to fundamental policy change. Hence, the third and final question of this study is:

What were the preconditions for large-scale tax reform?

In other words, this study poses one descriptive and two explanatory questions about the Norwegian tax reform of 1992. The first question concerns the outcome of the reform, whereas the two latter questions are about the process that led to this outcome.

Expressed more schematically, the first question is about determining the value of the tax reform on the dependent variable ‘outcomes of policy change’, which is either liberalization or rational updating of policy (or a combination of the two). The aim of the second question is to explore the linkages between the value on the dependent variable and the value ‘large-scale reform’ on the independent variable ‘process of policy change’. The third question takes us one step further backwards in the causal chain, looking to explain large-scale reform. This means both identifying which variables account for the reform and how these variables produced large-scale policy change. I look at structural, ideational and institutional explanatory variables. Figure 1.1 provides a simple illustration of the assumed connection between the major variables in this study.

This study has a meso-level focus, as it investigates policy evolution in one particular area, that of taxation. I understand the more or less coherent set of policies in this area as a policy

Process/scale of change

Outcomes of change Political

institutions

Structural factors

Ideational factors

Figure 1.1: A model for studying reform

(8)

regime². The tax system it treated as a whole, implying that the study is not limited to any specific part of tax policy or type of taxes, or to taxes collected at any particular level of government. This choice is justified by the ambition to examine the total outcomes of the tax system in terms of revenue-raising, redistribution and efficiency. Moreover, the different parts of tax policy are closely inter-connected, meaning that an analysis of one element in isolation can easily be misleading.

At issue in this study is the transformation of Norwegian tax policy since the 1980s. The study focuses on the 1992 tax reform, as this was the most significant policy change of the period.

To understand the background for the reform, we follow the development of tax policy back to the mid-1970s. My in-depth investigation of the political process concentrates on the years from the release of an important public commission report on taxation in 1989 until the passage of the reform in 1991. To shed light on the implications of the reform, the study also scans the developments in policy and outcomes from 1992 up until today.

1.2 Substantive motivation

The primary motivation for this study is substantive. The study springs out of an interest for the transformation of the Norwegian welfare state regime since the 1980s. Did Norway abandon its social democratic welfare state model during these decades, or was the model rather updated to the political and economic environment of a new era? This question has been widely debated, and some works have attempted to analyze the issue in its entire breadth. In particular, Dølvik (2007) offers a comprehensive analysis. Yet, somewhat surprisingly, one of the integral elements of the Norwegian welfare state regime has been neglected in both public and scholarly debate: the tax system.

Simply, our understanding of how Norwegian tax policy has changed in recent decades is very limited. There is a dearth of political and political-economic research on the character of the major Norwegian reforms of tax policy, of which the 1992 reform was the most important.

We do know that the 1992 tax reform virtually remade the entire tax system, but we are ignorant about the implications of these changes. Did the tax reform put an end to egalitarian social democratic tax policy and replace it with a liberal regime? This could be connected to

2 This understanding of a ‘policy regime’ rests on weaker assumptions about coherence and stability in policy than the notion of ‘institutional equilibrium’ often employed in rational choice scholarship.

(9)

less redistribution, a weaker capacity to raise revenue, or more lenient taxation of business and capital. Such an evolution could have threatened the financial (and political) basis of the welfare state, providing pressures for retrenchment.

Or did the reform rather represent a successful updating of social democratic tax policy to new political-economic realities? It is possible that fundamental policy changes were necessary for the tax system to better achieve its core objectives in an increasingly complex economy. If this were the case, the 1992 tax reform could be regarded as a prerequisite for the survival and return of the social democratic model. For instance, a reform that reinforced the capacity to raise revenue would be consonant with the maintenance or continued expansion of the welfare state.

Thus, determining the outcomes of the 1992 Norwegian tax reform is the first goal of this study, expressed in the first research question. What does the existing literature tell us on this point? In Norway, tax policy issues have been treated almost exclusively by economists and law experts. As a consequence, the existing knowledge about the 1992 tax reform is limited in its perspective. While its legal aspects and economic effects are well covered, its broader political-economic implications are not. In the absence of such research, superficial analyses have shaped the perceptions of the 1992 tax reform.

An oft-repeated view is that the tax reform produced a dramatic growth in capital incomes, which benefited the well off and thereby increased income inequality (e.g. Dølvik 2007:310).

However, this understanding of the 1992 tax reform is not only incomplete, it is also fallacious. Firstly, it focuses exclusively on the direct redistributive effect of the tax system, while ignoring tax policy’s arguably more important function of generating revenue for redistribution in the welfare state. Secondly, the purported explosion in capital incomes upon closer investigation turns out to be a statistical artefact (Fjærli and Aaberge 2003).

Neither does the broader literature on the Scandinavian welfare state regimes have much to offer on the issue of tax policy change. This comprehensive and highly prodigious literature has focused on transfers and services (e.g. Korpi 1983; Hatland et al. 2001; Kumlin and Rothstein 2005) and labor market questions (e.g. Barth et al. 2003; Kongshøj Madsen 2005), while according little attention to the role of tax policy.

(10)

Fortunately, there are a couple of excellent political-economic studies of tax reforms in Sweden (Steinmo 1993; 2002) and Denmark (Ganghof 2007). As the three Scandinavian countries implemented similar reforms at approximately the same time, these studies constitute a relevant frame of reference for my analysis of the Norwegian case. The findings from these studies have inspired my hypotheses about the outcome of the Norwegian tax reform.

Steinmo (1993:165,171) finds that the Swedish tax reform of 1991 involved a shift in attention from equity goals towards efficiency concerns, and implied a downward redistribution of the tax burden. That is, the tax reform in Sweden represented a rupture with the old social democratic model and a shift in a liberal direction. Yet, in a more recent analysis of Swedish tax policy, Steinmo (2002) has revised this view. He finds that although

“tax policy is most certainly adapting to the new political economic realities”, this has not meant “the end of redistribution” or the death of the Swedish welfare state (Steinmo 2002:841).

Steffen Ganghof presents a different analysis of recent tax reforms, based especially on the Danish case. He argues that reforms were primarily aimed at rationalization and upgrading:

“The tax reforms did not fundamentally change the weighting of different tax policy goals (efficiency and equity) but tried to better achieve both goals” (Ganghof 2007:1066). Drastic changes in tax rules did not reverse the existing goals of tax policy; they were rather “a precondition for defending the underlying substantive status quo” of tax policy (Ganghof 2007:1062).

These alternative analyses give rise to my two hypotheses about the outcome of the Norwegian tax reform: ‘liberalization’ and ‘rational updating’. Put very briefly, liberalization would imply greater efficiency at the expense of equality, while rational updating would mean a better ability to achieve goals of both efficiency and equality. These two hypotheses are discussed in greater detail in section 2.1. To preview my findings, this study finds evidence for the latter hypothesis. The investigation shows that the Norwegian tax reform of 1992 was characterized by rational updating rather than liberalization.

(11)

1.3 Theoretical motivation

Although this study first and foremost addresses a substantive issue, it is at the same time theoretically motivated. In different ways, the study seeks to contribute to three theoretical debates. The first is the general discussion about the relationship between processes and outcomes of change. The second is the debate about the impact of Nordic political institutions on reform. The third concerns the interaction between factors behind tax reform, and more generally, behind economic policy change. Here, I briefly sketch these debates and how I aim to contribute to them. The theory is discussed in greater detail in chapter 2.

Firstly, the argument that substantive outcomes of change (i.e. liberalization or the lack thereof) are associated with specific processes of change (i.e. incremental change or large- scale reform) is central in the literature on institutional change in advanced political economies³. What this relationship looks like is at the very core of current scholarly debate.

(For the full discussion, see section 2.2.).

Paul Pierson (2000) suggests that small, incremental policy change produces continuity in outcomes, while abrupt, large-scale change causes discontinuity in outcomes. Substantively, discontinuity in outcomes is often identified with liberalization or welfare state retrenchment, while continuity equals the lack thereof. Pierson’s view is challenged by Streeck and Thelen (2005), who argue that major discontinuities often result from incremental policy changes.

They contend that liberalization is usually the product of small, gradual changes in policy.

Conversely, one would expect that non-liberal reforms are most often the result of large changes in policy. Streeck and Thelen’s (2005) argument strongly suggests such a relationship, but they do not devote further attention to the type of policy change characterized by the combination of fundamental reform and continuity in outcomes.

However, the can place the Norwegian tax reform of 1992 into this category, as I find that this large-scale reform led to a rational updating of policy. The Norwegian case thus provides an opportunity to explore the relationship between fundamental policy change and continuity in

3 The notion of ‘institutions’ in this body of work encompasses ‘policy regimes’ as here defined. A policy regime can be regarded as a simple form of institution. The choice to conceptualize the Norwegian tax system narrowly, as a ‘policy regime’ rather than as an ‘institution’, reflects the particular character of tax policy. In the field of taxation, formal rules are all-important. Compared to other policy areas, tax policy can be identified more closely with tax rules. Therefore, we can avoid the more complex notion of ‘institution’. This also distinguishes policies from political institutions, discussed in section 2.4.

(12)

outcomes. The present study can potentially shed light on the mechanisms that underlie this relationship and thereby contribute to the development of theory on this point.

Secondly, the literature on institutional change in advanced political economies emphasizes the difficulty of large-scale policy change. This is generally attributed to the status quo-bias of political institutions. Paul Pierson (2000) argues that increasing returns processes prevalent in politics generate growing barriers to fundamental policy shifts. Why the Norwegian political system was able to produce large-scale tax reform is therefore of theoretical interest.

There is reason to suspect that the consensual Nordic political systems represent institutional settings that under some conditions are conducive to large-scale policy change (see section 2.4). Tranøy (2000) argues that in times of crisis, the consensus-oriented Norwegian system can be expected to produce fundamental updating reform. But exactly how do Nordic political institutions potentially stimulate reform? By exploring the institutional mechanisms that contributed to large-scale tax reform in Norway, I seek to contribute to this discussion.

Thirdly, there is theoretical debate about the factors that explain change, both in tax policy and in economic policy more generally. (For the full discussion, see section 2.2.) How important are institutional, structural and ideational factors, respectively, and how do they interact to bring about change? Traditionally, structural factors have been emphasized in explanations of economic policy regime change. Yet, lately, scholars that stress the importance of ideas as causal factors have challenged this materialist approach.

Mark Blyth (1997; 2001) argues that structural explanations are incomplete, since structure can only account for the breakdown of an old regime, not the character of the new regime that replaces it. Therefore, economic ideas can be seen as “a key mediating variable between structural change in the economic realm and institutional change in the political realm” (Blyth 2001:5). The present study addresses this general debate by investigating the interaction between structural, ideational and institutional factors in the process that led to the Norwegian tax reform.

At the same time, the study speaks to the specific debate about the factors behind recent tax reforms. In the existing literature on tax policy change, structure, ideas and institutions are all regarded as sources of tax reform. Ganghof (2007) stresses the importance of domestic

(13)

structural factors, arguing that the structural problems of extant tax regimes gave impulses to fundamental tax reform in the 1980s and 90s. Others emphasize international structural factors, arguing that economic globalization has imposed major structural constraints on tax policy (e.g. Genschel 2002; Bretschger and Hettich 2002). Steinmo (2003), on the other hand, highlights the role of new policy ideas in accounting for the major tax reforms of the 80s and 90s. Moreover, Steinmo’s (1993) seminal study of the evolution of tax policy in the USA, the UK and Sweden shows the deep impact of political institutions on tax policy change.

1.4 A historical institutionalist research strategy

The present study is based on a historical institutionalist (HI) strategy for framing research questions and developing explanations. This approach is usually juxtaposed to rational choice institutionalism. Basically, what distinguishes the HI strategy from rational choice modelling is the focus on large, substantive questions, as well as the construction of explanations that are sensitive to context and based on the combined effects of institutions and processes.

Firstly, historical institutionalists address big, substantive research questions rather than theoretical questions (Pierson and Skocpol 2002:3). “Historical institutionalists are primarily interested in understanding and explaining specific real world political outcomes,” as Steinmo (2001) puts it. Accordingly, the main goal of the present study is substantive, namely to understand and account for major changes in Norwegian tax policy. HI research is inherently problem driven, in contrast to rational choice scholarship, which is usually more theory driven (cf. Shapiro 2005).

Secondly, to explain political outcomes, historical institutionalists investigate the “combined effects of institutions and processes” in historical context (Pierson and Skocpol 2002:3;

emphasis in original). The present study analyzes how different structural and ideational processes interacted with political institutions during a specific historical period to bring about tax reform. This approach to explanation differs from rational choice modelling, which will often investigate the effect of one specific institution and accord less attention to contextual conditions. Historical institutionalists usually understand institutions as intervening or structuring variables among a large set of explanatory factors, rather than as the sole variable that has impact (Steinmo 2001).

(14)

A third feature that distinguishes HI from rational choice institutionalism is the perspective on preference formation (Thelen and Steinmo 1992). Historical institutionalists treat preferences as endogenous, whereas rational choice modelling has often assumed exogenous preferences.

This difference is not as clear as before, as the issue of preferences is at the top of the agenda for theory development within rational choice scholarship. Yet, it is fair to say that HI scholars treat preferences as endogenous in a more fundamental sense: “[T]he definition of interests and objectives is created in institutional contexts and is not separable from them”

(Zysman 1994:244). The present study adopts this understanding. One of the core issues of the analysis is how the preferences of political actors were shaped within the Norwegian institutional setting.

However, we should not exaggerate the differences between HI and rational choice institutionalism. For one, historical institutionalism is not incompatible with rational action.

HI research – the present study included – is based on the assumption that “most people act rationally most of the time” (Steinmo 2001). Importantly, this refers to rationality in the broad sense of the word, not a notion of rationality confined to economic self-interest.

For another, many HI studies incorporate elements from rational choice theory. Most importantly, “historical institutionalists have … taken on board the notion that institutions that solve collective action problems are particularly important in understanding political outcomes” (Thelen 1999:370). The notion of political institutions as systems that can generate cooperation is at the core of the present study. To explain how institutions contributed to collective political action, I borrow certain concepts and insights from more instrumental analyses.

1.5 Research design and methodology

The investigation of my research questions is designed as a case study. The case study design is appropriate, as it allows for in-depth investigation of both the outcomes of the 1992 Norwegian tax reform and the process behind it. Understanding this specific case is the first part of the case study’s double ambition (cf. Gerring 2004). The other part is to draw general inferences from this case to a broader class of cases. But what is the Norwegian tax reform of 1992 a case of? This question does not have one simple answer. Since my three research

(15)

questions refer to partly different bodies of theoretical literature, they give rise to different specifications of the broader class of cases.

First, the Norwegian tax reform is a case of modern tax policy change, in particular in the Nordic countries, when it comes to understanding and explaining policy change (questions 1 and 3). Second, it is a case of a broader theoretical category, namely policy regime change in advanced political economies, when the link between processes and outcomes of change is at issue (question 2). This is also the potential scope for arguments about the interaction between structural, ideational and institutional factors behind policy change (question 3). Third, the Norwegian tax reform provides a case of how Nordic political institutions affect policy change (question 3).

Consonant with the historical institutionalist research strategy, this study takes a mechanism- based approach to explanation (cf. Pierson and Skocpol 2002:6). The goal is to explore the causal mechanisms that connect initial conditions to outcomes, not to estimate causal effects.

This approach “seeks to explain a given social phenomenon … by identifying the processes through which it is generated” (Mayntz 2004:238).

The ambition to uncover mechanisms goes hand in hand with the methodology of process tracing. Process tracing implies close-up investigation of every link in the chain of events that led to the large-scale 1992 tax reform. This is crucial to examine the combined effects of institutions and historical processes.

Specifically, this study relies on three main sources of data: qualitative interviews with policy-makers and experts; public documents; and secondary statistical economic literature.

While qualitative interviews and public documents are used to trace the reform process, existing statistical economic studies are the main basis for the discussion of the outcome of reform. The research design and methodology of this study is discussed in detail in chapter 3.

1.6 The importance of tax policy

Finally, a brief introduction to the substance of tax policy is necessary to prepare the reader for the theoretical and empirical discussions that follow. Why is tax policy important, and what are its main functions?

(16)

The tax system represents an integral element of a country’s welfare state regime or welfare capitalism. The basic purpose of taxation is to finance the activities of the public sector. In advanced political economies, particularly in the Scandinavian countries, a large portion of public spending goes to welfare services and transfers. In 2008, social protection and health care accounted for 56,7 percent of Norwegian public spending, according to Statistics Norway. Simply put, taxes pay for the welfare state. Although this statement is fairly trivial, its implications are crucial for understanding the importance of tax policy.

First, the capacity to raise tax revenue constrains the size of the welfare state. Since the tax system represents the income side of the welfare state, it also constitutes its budget constraint.

The scope for welfare state spending is limited by two features of the tax system: first, how effective it is in terms of raising revenue, and second, how much taxes people are willing to pay. In other words, the government’s opportunity to conduct welfare policy will increase with both the effectiveness and the popular acceptance of tax policy.

This is the case even for Norway, which benefits from large petroleum incomes. Taxes are necessary to make room in the real economy for the activities of the public sector. Through taxation, resources – such as human capital – are transferred from the private sector to the public sector.

Second, taxation influences economic efficiency. Basic economic theory says that taxes entail efficiency losses, since they distort the incentives of firms and individuals⁴. More interestingly, the size of the efficiency loss varies with the type of taxation, since different taxes affect the economy in different ways. Therefore, how we tax determines the economic costs associated with raising revenue through taxation. The tax costs for the economy depend both on the chosen tax mix – i.e. the combination of different types of taxes – and on the specific design of each tax.

4 Taxes that correct externalities constitute an exception, as they are meant to prevent efficiency losses by confronting firms or individuals with the social economic costs and benefits of production or consumption.

(17)

An important implication is that the economic cost of the welfare state⁵ is partly determined by how the tax system is designed (Lindert 2004). A more efficient tax system makes the welfare state less costly in economic terms.

More broadly, tax policy strongly influences the working of the economy. Taxes affect the economic decisions of individuals and businesses. For individuals, taxes influence the choice between work and leisure, between consumption and savings, and between different forms of savings. While one tax system may give people incentives to work and put the money in the bank, another system may induce individuals to borrow money and buy real estate. For businesses, taxes affect both the decision of whether to invest or not, and the choices between different types of investments and between alternative ways to finance investments. One tax system may deter investments, while another may favor loan-financed investments.

Third, tax policy has a redistributive function. Welfare policy and tax policy are alternative arenas for redistribution, since they constitute respectively the spending side and the income side of the welfare state. While benefits are distributed through welfare policy, tax policy distributes burdens. To achieve a desired level of redistribution, policy-makers can thus choose different combinations of welfare policy and tax policy. For instance, minimal redistribution in the tax system can be compensated by strongly redistributive welfare policies.

Taxes have different effects in terms of redistribution, as they can be either proportional, progressive or regressive. Taxes are proportional if they represent the same proportion of income at all income levels. They are progressive if they take away a higher percentage of income from high-income earners than from low-income earners. Conversely, regressive taxes represent a larger proportion of income for those at low income levels compared to those at high levels of income.

Importantly, the total redistributive effect of the tax system depends both on the chosen mix of taxes and the specific design of these taxes. Some types of taxes are generally more redistributive than others. Wealth taxes and regular income taxes are usually progressive,

5 The term “cost of the welfare state” does not imply that the costs associated with the welfare state are greater than its benefits. It simply refers to the fact that the welfare state mainly has to be financed through taxes that distort economic choices to a smaller or larger degree. Taxes that are less distortionary will lower the total tax costs associated with the welfare state.

(18)

while consumption taxes and payroll taxes are often proportional or even regressive. For instance, a flat consumption tax will normally be de facto regressive, since low-income earners use a larger portion of their income on consumption than high-income earners. The specific structure of each tax also has distributive implications. Tax ceilings generally make taxes more regressive, while tax floors or standard deductions make them more progressive.

Additionally, so-called ‘tax expenditures’ have considerable effects on redistribution (Ervik 2000). Tax expenditures refer to the forgone revenues from taxing some assets more leniently than what is the norm in the tax system. These ‘expenditures’ normally take the form of deductions or low value assessments of assets, which may have highly asymmetrical effects.

For instance, lenient taxation of owner-occupied housing privileges those with (large) houses.

In sum, how tax policy is designed is important for three main reasons. First, it determines the revenue-raising capacity of the tax system, which constrains the size of the welfare state.

Second, it influences the working of the economy and partly determines the economic costs of the welfare state. Third, it affects the distribution of income.

1.7 Outline of the study

The study is structured as follows: In chapter 2, I present hypotheses, theory and relevant existing literature. This chapter is divided into five sections. In the first section, I present the two hypotheses about the character of reform, liberalization and rational updating. In the second section, I sketch the theoretical discussion about the relation between processes/scale and outcomes of institutional change in advanced political economies. Section 2.3 discusses the existing literature on structural and ideational sources of tax policy change. In section 2.4, I address the workings of political institutions in the Nordic countries. Finally, I discuss the roles of key political actors operating within this institutional context (section 2.5).

The third chapter is devoted to research design and methodology. In chapters 4 and 5, I present the results from my empirical investigation. The process that led to reform is the subject of chapter 4, while chapter 5 is about the outcome of reform. Chapter 6 contains the analysis, where I bring theory and data together to answer my research questions. Chapter 7 concludes and suggests topics for further research.

(19)

Chapter 2: Theory

2.1 Liberalization or rational updating?

2.1.1 Two hypotheses about the outcome of the 1992 tax reform

The first research question is descriptive and addresses the character of tax policy change: Did the Norwegian tax reform of 1992 imply a liberal policy shift or a rational updating of social democratic policy? In the literature on recent tax policy change there are two main views of the reforms of this era. The first is that tax policy was subject to liberalization, with attention shifting from equity to efficiency goals. The alternative view is that tax policy was rationally updated, as policy-makers found more effective means to achieve a fixed set of goals. These views are the basis for my two hypotheses about the Norwegian tax reform of 1992, since they suggest the two different paths tax policy could have taken at this crossroads.

The hypotheses, which I refer to as ‘liberalization’ and ‘rational updating’, are also tied to a more general discussion in the literature on policy regime change in advanced political economies. In this section, I develop the two hypotheses, describing both their general features and their specific implications for tax policy in Norway.

2.1.2 Liberalization

The first and most conventional hypothesis is that of liberalization. Liberalization is regarded as a dominant contemporary trend both in works on tax policy change and in the broader literature on changes in advanced political economies. Generally, liberalization is often identified as a shift to a more market-based political-economic logic that involves larger freedom for private economic actors (e.g. Streeck and Thelen 2005).

In the tax field, the discussion about liberalization is more oriented towards the balance between policy objectives of efficiency and equality, as well as the corresponding outcomes.

In the tax reforms of the 1980s and 90s, some analysts see a “shift in policy paradigm” where the goal of efficiency was given increasing weight relative to that of equity (Swank and Steinmo 2002:643,651). Steinmo finds “an unmistakably common trend: Tax reform has now

(20)

come to mean the redistribution of existing tax burdens downward” (Steinmo 1993:156). That is, policy change in a liberal direction is identified with a shift from equity towards efficiency.

Accordingly, I conceptualize ‘liberalization’ as a shift in policy objectives from equality towards efficiency and a corresponding development in outcomes. Liberalization implies greater efficiency at the expense of equality. Thus, I regard a weakened ability to generate economic equality as a necessary precondition for classifying a development as liberal.

The proposed “big trade-off” (Okun 1975) between efficiency and equity constitutes the basis for this notion of liberalization. This basic tenet of neo-classical economic thought implies that stronger redistribution leads to lower efficiency, while higher efficiency means less redistribution. In this trade-off, liberalization can be equated with giving priority to the goal of efficiency over that of equality: Redistributive efforts are problematic since they hurt efficiency, and should therefore be reduced.

More generally, this neo-classical view implies that government intervention should be limited, since it distorts the working of free, competitive markets, thereby impeding an optimal allocation of resources. Liberalization thus involves loosening the grip on private economic actors, so that they can make free economic choices that generate economic efficiency.

What are the specific implications of this hypothesis for tax policy in Norway? On the policy regime level, the liberalization hypothesis corresponds to a shift from a traditional interventionist social democratic tax policy model to a liberal regime.

Liberalization would change the weight accorded to different objectives of tax policy. On the one hand, liberalization implies a stronger concern for efficiency and the free operation of markets. More lenient tax treatment of business and mobile capital would be the logical consequence. Thus, the outcomes related to liberalization would be greater economic efficiency and a lighter tax burden on business and capital.

On the other hand, a liberal tax policy regime would accord less weight to goals of redistribution. This applies both to the direct redistribution within the tax system and the indirect redistribution based on raising revenue for the welfare state. Thus, liberalization

(21)

would be related to outcomes of higher inequality, either because of a downward shift of the tax burden or a weakened capacity to finance redistributive welfare policy.

To summarize, the first hypothesis is that the 1992 Norwegian tax reform changed policy in a liberal direction. In terms of policy objectives, this would imply a shift from equity to efficiency concerns. In terms of outcomes, liberalization would be associated with higher economic efficiency and a lower tax burden on business/capital, combined with greater inequality and a reduced ability to finance the welfare state.

2.1.3 Rational updating

The alternative hypothesis about how Norwegian tax policy developed is what I label rational updating. This hypothesis has two basic elements. ‘Rational’ refers to using more rational policy means to achieve a fixed set of goals. ‘Updating’ means adapting policy to new political-economic environments. Hence, rational updating of policy means to better achieve established objectives by adapting policy means to new political-economic realities.

Though rarely stated explicitly, this hypothesis is anchored both in the general literature on changes in advanced political economies and in discussions of tax policy change. On a general level, Hacker (2005) and Streeck and Thelen (2005) provide a theoretical foundation for rational updating. The latter pair write that “institutions require active maintenance; to remain what they are they need to be reset and refocused, or sometimes more fundamentally recalibrated and renegotiated, in response to changes in the political and economic environment” (Streeck and Thelen 2005:24). The point is that policy changes are often necessary to better achieve the core objectives of a policy regime. Policy changes can be of a predominantly rational character, and do not necessarily imply changes in the substance of a policy regime.

Paul Pierson (2001) also writes about rational reforms in his discussion of welfare state restructuring. He mentions ‘rationalization’ as one of two reform types under the general dimension of ‘recalibration’. Pierson defines rationalization as the “modification of programmes in line with new ideas about how to achieve established goals” (Pierson 2001:425).

(22)

In the tax field, Steffen Ganghof’s analyses of modern tax reforms strongly suggest the idea of rational updating. He argues that reforms were primarily aimed at rationalization and upgrading: “The tax reforms did not fundamentally change the weighting of different tax policy goals (efficiency and equity) but tried to better achieve both goals” (Ganghof 2007:1066). Policy changes were necessary to ensure continuity in outcomes. Drastic changes in tax rules did not reverse the existing goals of tax policy; they were rather “a precondition for defending the underlying substantive status quo” of tax policy (Ganghof 2007:1062).

Steinmo’s recent analysis of Swedish tax policy also partly suggests the hypothesis of rational updating. His point is that although “tax policy is most certainly adapting to the new political economic realities”, this has not meant “the end of redistribution” or the death of the Swedish welfare state (Steinmo 2002:841).

Whereas the hypothesis of liberalization implies a shift in objectives, rational updating entails continuity in objectives. Existing goals are not reversed or altered; only the policy means for achieving these goals change. Contrary to the first hypothesis, rational updating does not assume that there is a trade-off between efficiency and equality. The notion of rational updating implies that it is possible to better achieve both goals simultaneously, or at least to increase efficiency without hurting equality, and vice versa.

Identifying ‘rational’ policy change is a challenging exercise analytically. For one, the term

‘rational updating’ is defined by stability regarding objectives and amelioration in terms of achieving them. Can policy changes really be wholly neutral and rational? My answer would be that although changes can never be completely neutral, in some cases the continuity in objectives/outcomes will be more significant than the discontinuity. A notion of rational policy change is warranted, as we clearly do not want to identify all modern changes in economic policy as liberalization (cf. Pierson 2001:425).

Yet, wouldn’t all policy-makers claim that their policies are rational, since they aim to better achieve goals? To meet this challenge, it is essential to base the evaluation of policy change primarily on outcomes. This study undertakes a thorough investigation of the outcomes of policy change to determine whether liberalization or rational updating characterized the 1992 tax reform.

(23)

What are the specific implications of the rational updating hypothesis for tax policy in Norway? On the policy regime level, this hypothesis corresponds to replacing the traditional social democratic model with a modern social democratic tax policy regime.

Rational updating implies continuity in the core objectives of tax policy. Such policy change would thus reconcile the goals of efficiency, revenue-raising and redistribution. Efficiency in taxation would not be pursued at the expense of equity. The outcome of rational updating for the tax system would be both greater economic efficiency and a greater total redistributive capacity. Increased redistributive capacity means a strengthened ability to finance the welfare state and/or greater direct redistribution within the tax system.

To sum up, the second hypothesis is that Norwegian tax policy moved in a rational direction in the 1992 reform. This is associated with continuity in objectives. In terms of outcomes, this would imply greater efficiency combined with more effective revenue-raising and stability in the distribution of tax burdens. Table 2.1 summarizes the implications of the two hypotheses for changes in the outcomes of tax policy.

Table 2.1: Implications of the hypotheses for outcomes of tax policy change

Hypothesis

Outcome variable Liberalization Rational updating

1. Efficiency Greater Greater

2. Treatment of capital More lenient Same as before or tougher 3. Capacity to finance welfare state Weaker Same as before or stronger

4. Direct redistribution Weaker Same as before or stronger

2.2 Linking outcomes and processes of change

2.2.1 Institutional change in advanced political economies

The second research question of this study – How was the outcome of the 1992 tax reform conditioned by the large scale of policy change? – is inspired by the literature on institutional change in advanced political economies. Substantively, this body of work has addressed the character and outcomes of the major changes that have taken place in various policy areas

(24)

and broader institutions since the 1980s. Theoretically, it has been concerned with understanding processes of change. However, it is the link between the two that constitutes the decisive point at issue in this literature. The argument that substantial outcomes are associated with specific types of institutional change represents the very raison d’être of this line of theoretical effort.

Most scholars of institutional change support the proposition that there is a relationship between outcomes and processes of change. What this relationship looks like, however, is the subject of intense academic debate. The discussion about what type of institutional change gives rise to specific substantive outcomes is arguably the most important current debate in the field. There are two major positions in this discussion: the first associated with Paul Pierson’s work; the other most clearly represented by Kathleen Thelen and colleagues. I first present these two positions, before I go on to argue that the kind of change I find in Norwegian tax policy represents a related but distinct theoretical category that has been given little attention in this literature.

2.2.2 Path dependence and the lack of retrenchment

The first position is based on the work of Paul Pierson and others on welfare state retrenchment. The substantial question posed by the “retrenchment literature” is whether the welfare states of advanced democracies have been dismantled as a consequence of increasing strains. The main finding is that the welfare state has proven remarkably resilient to change.

Although there have been cuts, most systems of social protection have resisted fundamental shifts. The result of increasing pressures has not been across-the-board retrenchment or liberalization, but rather limited restructuring of the advanced welfare state (Pierson 2001).

Thus, the substantive outcome has been continuity rather than change.

The retrenchment literature is, however, equally concerned with processes of change as with outcomes. Regarding processes, the dominant finding is that changes have been incremental, not abrupt. Taken together, “most reforms in most countries [have been] incremental rather than radical, and focused on restructuring rather than straightforward dismantling” as Pierson sums up an anthology on changes in advanced welfare states (Pierson 2001:420).

(25)

Paul Pierson makes the conceptual link between these two findings, arguing that the outcome of limited retrenchment is associated with a specific process of institutional change, namely incremental, path dependent change. Welfare state institutions have not been dismantled, because they tend to evolve according to a path dependent logic. As institutions develop along a particular path, processes of increasing returns are set into motion, which make it increasingly difficult to exist the existing path. These processes produce significant barriers to fundamental policy shifts. Change, therefore, will mostly be incremental, amounting only to adjustments and adaptations within the existing path.

According to this view, the process of change thus accounts for the outcome of change, namely institutional stability. Simply put, incremental processes of change produce outcomes of institutional continuity. Arguments about path dependence account for institutional persistence.

However, the outcome to be explained is not always institutional stability. Major institutional changes do indeed occur. The path dependence literature usually points to the concept of

“critical junctures” to make sense of fundamental shifts of this kind. Critical junctures are the moments of creation or innovation “where the usual structural constraints on action are lifted or eased” (Mahoney and Thelen 2009), often caused by large exogenous shocks. During such critical moments, radical shifts from one institutional path to another are possible.

A common theoretical view is that institutional development over long periods of time incorporates both the logics of path dependence and critical junctures. This is often called the punctuated equilibrium model. Institutional change is characterized by long path dependent continuities that are periodically interrupted by critical moments when radical shifts occur (Pempel 1998). In the stretches of time between these rare critical junctures, institutions develop in path dependent ways.

Arguments about path dependence and critical junctures should not be conflated, as stressed by Pierson (2000). The two concepts point to diametrically opposite logics of change. Path dependence arguments emphasize the barriers to change that explain stability, while arguments about critical junctures stress the lifting of these barriers, which accounts for major ruptures. However, they give rise to complementary hypotheses about the relationship

(26)

between processes and outcomes of change. While incremental path dependent change produces institutional continuity, abrupt change (at critical junctures) causes discontinuity.

2.2.3 Liberalization through gradual transformation

Wolfgang Streeck and Kathleen Thelen criticize this punctuated equilibrium model, arguing that there are “severe limits to models of change that draw a sharp line between institutional stability and institutional change” (Streeck and Thelen 2005:8, applies also to next quotes).

Separating the analysis of institutional continuity and rupture, they argue, imposes a bias on our understanding of the relationship between processes and outcomes of change. This conceptual framework only provides for “either incremental change supporting institutional continuity through reproductive adaptation, or disruptive change causing institutional breakdown and innovation and thereby resulting in discontinuity”.

Equating incremental change with continuity and abrupt change with discontinuity is misleading, they argue, since discontinuities can also arise from incremental changes: “Far- reaching change can be accomplished through the accumulation of small, often seemingly insignificant adjustments”. To account for this possibility, they suggest that we distinguish clearly between “processes of change, which may be incremental or abrupt, and results of change, which may amount to either continuity or discontinuity”. This is presented in a two- by-two matrix, reproduced under (Table 2.2).

Table 2.2: Types of institutional change: processes and results

Result of change

Process of change Continuity Discontinuity

Incremental change 1. Reproduction by adaptation

3. Gradual transformation Abrupt change 2. Survival and return 4. Breakdown and

replacement Source: Streeck and Thelen (2005:9)

This typology of institutional change maps out four possible combinations of processes and results of change. But the typology also represents four possible connections between the

(27)

character of the change process and the outcome of change. Cell 1 represents Pierson’s view, which conceives of incremental change as adaptive and reproductive, thus implying continuity. Cell 4 corresponds to arguments about critical junctures, where abrupt change causes institutional rupture. Taken together, the punctuated equilibrium model allows for institutional change of types 1 and 4.

Streeck and Thelen, however, emphasize institutional change of type 3, which they term

‘gradual transformation’. “In reality,” they argue, there can be “dramatic institutional reconfiguration beneath the surface of apparent stability or adaptive self-reproduction, as a result of an accumulation over longer periods of time of subtle incremental changes” (Streeck and Thelen 2005:8). They regard this as a frequent – maybe even the predominant – type of institutional change in the political economy of modern capitalism. But due to the dominance of the punctuated equilibrium model, these processes of gradual transformation are poorly understood.

Their basic point is that institutional discontinuity is not necessarily the result of abrupt change brought about by exogenous shocks. On the contrary, major transformations of institutions are most often produced by series of small, seemingly insignificant changes. This represents an alternative view of the relationship between processes and outcomes of change:

Major discontinuities result from incremental change.

In the context of modern political economies, discontinuity is inextricably linked to liberalization. Streeck and Thelen’s argument is based on the observation that advanced industrial democracies have gone through a process of liberalization since the 1980s that has fundamentally altered their political economies. They thus present a rival view of the substantive outcome of the pressures put on advanced political economies. Contrary to the

‘retrenchment’ and ‘Varieties of Capitalism’ literatures, which emphasize continuity and the lack of convergence, they stress the significance of the changes that have taken place. They see “a process of liberalization [that] involves a major recasting of the system of democratic capitalism as we know it” (Streeck and Thelen 2005:5).

Liberalization, however, has not been produced by disruptive changes caused by exogenous shocks. Quite the opposite: “[A]n essential and defining characteristic of the ongoing worldwide liberalization of advanced political economies is that it evolves in the form of

(28)

gradual change that takes place within, and is conditioned and constrained by, the very same postwar institutions that it is reforming or even dissolving” (Streeck and Thelen 2005:4).

The authors thus observe a relationship between the outcome of liberalization and incremental processes of institutional change. These processes of gradual but transformative change are of four kinds (Mahoney and Thelen 2009:15-17). First, displacement occurs when old rules are replaced with new ones, for instance when private institutions are introduced that compete with and eventually supplant public schemes. Second, layering implies that new rules are added to existing rules, thereby altering the logic of an institution. Third, drift describes processes where shifts in the context of policies alter the effects of policies that are not updated (Hacker 2005). That is, even if policies are not changed, changing environments can bring about significant changes in the outcomes of policies. For instance, public welfare schemes may decay because they are not tended to. Fourth, conversion occurs when existing rules are interpreted or enacted in new ways. Often, actors exploit the ambiguities of institutional rules, filling institutions with new goals or functions.

Streeck and Thelen argue that gradual transformation processes of these kinds abound in the recent development of advanced political economies. This brings them to ask whether the connection between incremental processes of change and outcomes of liberalization is of a more general character: “Could it be that measures of liberalization are somehow particularly suited to being imposed gradually and without disruption?” (Streeck and Thelen 2005:33).

They support this suspicion by pointing out that liberalization seldom requires political mobilization, as the typology of gradual transformative change suggests. Liberalization can be brought about without dismantling existing institutions, simply by doing nothing or by changing institutions from within.

Streeck and Thelen thereby provide an argument about the linkage between processes and outcomes of institutional change that runs counter to Pierson’s understanding. While Pierson’s view is that incremental change causes continuity and lack of retrenchment, they argue that gradual change privileges discontinuity in the form of liberalization.

(29)

2.2.4 Fundamental changes, continuity in outcomes

The possible linkages between processes and outcomes of institutional change are, however, not exhausted. Returning to the typology of institutional change in table 1, the punctuated equilibrium model corresponds to cells 1 and 4, while the model of gradual transformation occupies cell 3. What about the combination of abrupt change and continuity in outcomes (cell 2)? Streeck and Thelen note that “there often is considerable continuity through and in spite of historical break points”, tentatively referring to this type of change as ‘survival and return’ (Streeck and Thelen 2005:8). This category of institutional change is not given further attention, though.

However, the kind of change that I find in the Norwegian tax reform of 1992 seems to fit this category. My empirical investigation reveals a remarkable stability in outcomes, despite – or perhaps because of – abrupt policy change. This type of institutional change appears compatible with the argument advanced by Streeck and Thelen. The criticism of equating processes of change with outcomes applies here as well: Neither does incremental change signify continuity in outcomes, nor does abrupt change imply discontinuity. Their claim is that liberalization often results from the gradual change generated by the failure to adapt policy to changing environments (see also Hacker 2005):

“[I]nstitutions require active maintenance; to remain what they are they need to be reset and refocused, or sometimes more fundamentally recalibrated and renegotiated, in response to changes in the political and economic environment in which they are embedded […] Without such tending … they can be subject to erosion or atrophy through drift” (Streeck and Thelen 2005:24, emphasis in original)

This argument, which underpins the claim about ‘gradual, but transformative’ change, also implicitly supports the type of institutional change characterized by continuity in outcomes through disruptive large-scale reform. In this view, liberalization often occurs due to the lack of major reform that updates an institution to a new economic or social environment. The logical implication is that major, updating reforms can indeed produce institutional continuity and deter liberalization. Ganghof advances the same point with regard to tax policy, arguing that “drastic changes in the legal status quo (policies) may be necessary to defend the underlying substantive status quo (outcomes)” (Ganghof 2007:1081).

(30)

This argument can be pushed even further: Could there be a connection between non- liberalization and abrupt, fundamental reform? Again, this suspicion mirrors Streeck and Thelen’s speculation about a general relationship between gradual change and liberalization.

While liberalization can easily proceed without mobilization, updating institutions to changing contexts is a more difficult task that requires mobilization. That is, liberalization does not depend on large reforms, while non-liberalization by means of recalibrating institutions does.

“Non-liberal reforms in a market economy seem to require ‘political moments’ in which strong governments create and enforce rules that individual actors have to follow, even if they would on their own prefer not do so”, Streeck and Thelen (2005:33) write. As the authors suggest, organizing capitalism today may face far stronger collective action problems than liberalization within capitalism.

However, their emphasis on ‘strong government’ as the only solution to these problems seems misleading. Broad, political coalitions behind reform represent a strong – and perhaps more viable – alternative. That is, collective political action by broad coalitions appears to be the most important condition for fundamental non-liberal reform.

2.2.5 Summary of the discussion of outcomes and processes of change What is the relationship between processes of institutional change and substantial outcomes in advanced political economies? In the preceding discussion, I have presented two conceptual models, which each give rise to two types of institutional change. The punctuated equilibrium model suggests that incremental changes imply continuity in outcomes, while abrupt changes produce discontinuity. Pierson explains lack of retrenchment with the path dependent character of institutional change (‘reproduction by adaptation’). Major institutional discontinuities (‘breakdown and replacement’), on the other hand, are understood as the product of disruptive change at critical junctures.

Streeck and Thelen present an alternative model, arguing that the connection between processes and outcomes of institutional change has the opposite sign. Discontinuity in the form of liberalization most often occurs through incremental changes (‘gradual

Social democratic tax policy updated : A study of the Norwegian tax reform of 1992